Multiplex ligation-dependent probe amplification (MLPA)

MLPA was conducted according to the supplied manufacturer’s specifications using the probemix P022-PLP1 (MRC-holland, Amsterdam, Holland). PCR products were then mixed with formamide (HiDi Formamide, Applied Biosystems, Foster City, CA) and fluorescent Genescan 500 LIZ size standard (Applied Biosystems) prior to analysis using an Applied Biosystems ABI 3130xl capillary sequencer (Applied Biosystems). Data collection and export used GeneMapper software (Applied Biosystems). Peak areas of each fragment were compared to those of a control sample to calculate the gene dosage of each amplicon, including those corresponding to each exon of the PLP1 gene.

X chromosome inactivation pattern

The analysis of the X inactivation pattern was based on extend human androgen receptor (HUMARA) assay proposed Bertelsen et al. [8]. Briefly, 200 ng of leukocyte DNA was digested with HpaII (New England Biolabs, Ipswich, MA) at 37 °C for 12 h. PCR amplification of four loci (AR, PCSK1N, SLITRK4, and ZDHHC15) was performed on undigested DNA and HpaII digested DNA. PCR products were subsequently analyzed on an ABI 3130xl Genetic Analyzer (Applied Biosystems) using GeneMapper software (Applied Biosystems). The X-inactivation ratios were calculated as previously described and skewed X chromosome inactivation was considered if this ratio decrease below 20 % in the investigated blood sample [8].

PLP1 gene expression analysis in patient fibroblasts

Total RNA was isolated from fibroblasts of eight controls without copy number variation of the PLP1 gene, of one positive control with a duplication of PLP1, and of our patient using the guanidinium thiocyanate-phenol-chloroform method. Ten microgram was used to generate cDNA using Superscript II first-strand cDNA synthesis kit (Invitrogen, Carlsbad, CA, USA) with oligo (dT) primers, according to the manufacturer’s protocol. Our patient and control cDNAs were used for PLP1 gene expression analysis performed by real-time PCR according to previously reported conditions [9]. Specifically designed primers were used to determined total PLP1 gene expression after normalization to the beta-glucuronidase gene (GUSB) expression using the ΔΔCt method.

Microarray

Creation of a custom array CGH design was performed using the Agilent software eArray (https://earray.chem.agilent.com/earray/), referred to the GRCh37/hg19 assembly to cover the PLP1 genomic region with maximum resolution. The microarray contained a total of 30,032 probes spanning a 20 Mb region surrounding the PLP1 gene (Agilent Technologies, Santa Clara, CA). Those 30,032 probes were distributed between three groups with various probe densities. The first group includes 15,032 probes encompassing 2 Mb (chrX: 102,000,000–104,000,000; corresponding to a 150 bp average probe spacing) around the PLP1 gene and includes most of the PLP1 duplications described so far [10, 11]. A second, largest interval contained 10,000 probes surrounding 8 Mb (chrX: 98,000,000–102,000,000 and chrX: 104,000,000–106,000,000; 800 bp average probe spacing). The last interval was designed to encompass the largest PLP1 rearrangement and includes 5,000 probes spanning 10 Mb (chrX: 93,000,000–98,000,000 and chrX: 106,000,000–113,000,000; 2000 bp average probe spacing) [6].

A HumanCytoSNP-12 (V2.1) pangenomic microarray (Illumina, San Diego, CA USA) was also performed.

Fluorescence in situ hybridization (FISH)

Fluorescence in situ hybridization (FISH) was performed on lymphocytes metaphase spreads according to standard protocols. FISH analysis was carried out with the probe RP11-832L2 spanning 183 Kb (chrX:102,902,650–103,085,315), which encompasses the entire PLP1 gene, together with a subtelomeric probe of the short arm of chromosome 1 (CEB108/T7).

Genome walker

The Universal Genome Walking Kit (Clontech, Palo Alto, CA, USA) was used to identify precise breakpoints of the inserted PLP1-containing X chromosome segment into 1p36. According to the manufacturer’s protocol, the genomic DNA of the patient was completely digested with four separate restriction enzymes (i.e. EcoRV, DraI, PvuII, and SspI) and subsequently ligated to the GenomeWalker adaptors. The GenomeWalker libraries produced were then used as a template for primary PCR amplification. The primers used were the outer adaptor primer provided in the kit and a Xq22 specific primer localized at the boundaries of the duplicated PLP1 segment identified with our highly resolutive PLP1-custom array CGH. A secondary nested PCR using the nested adaptor primer and a nested Xq22 specific primer was then performed. Sequencing of the PCR products was finally achieved if a single or two (considering that both, the normal Xq22 and the 1p36 derivative, alleles could be digested and ligated to the adaptor) specific and major PCR products were visible after electrophoresis through an agarose gel stained with GelRed (Biotium, Hayward, CA, USA). Sequencing was conducted using the nested primers with the BigDye terminator sequencing kit (Applied Biosystems).

Case report

Here we report the case of a girl who is the third child from unrelated and healthy parents of Greek origin. Pregnancy was uneventful, however during the second month of life a horizontal nystagmus and severe hypotonia were noticed. At 9 months of age, a brain magnetic resonance imaging (MRI) was performed due to a persistent delay in motor milestones associated with ataxia of the head and trunk led; MRI revealed a severe delay in the myelination process (evaluated at 2 months of age). Afterwards, she improved obviously in motor capacities (holding her head up at 12 months, sitting with support at 16 months, and standing position at 30 months) with excellent psychosocial interactions and decreased nystagmus intensity. Another brain MRI at 33 months of age confirmed the diffuse “hypomyelinating” pattern of the supratentorial cerebral white matter (Fig. 1). In addition, brainstem auditory evoked potentials showed an exclusive cochlear wave I without subsequent recordable brain waves suggesting a severe impairment in CNS conduction. No other abnormalities including dysmorphic features, heart, kidney, or liver problems were found. Now aged 4 years, she continues to improve (crawling, walking with support, feeding herself, drawing a line, speaking using comprehensive simple sentences). However, signs of pyramidal tract dysfunctions are clearly present in the lower limbs and MRI abnormalities have not changed. Those clinical features strongly suggested the diagnosis of PMD.

Characterization of the PLP1 gene duplication

MLPA analysis of the DNA extracted from the patient’s blood revealed a duplication of the seven probes localized in each coding exon of the PLP1 gene. Additional probes that were duplicated revealed that the duplication extended at least from the TCEAL1 to ESX1 genes thus encompassing at least 614 Kb (but less than 1685 Kb).

To confirm the PLP1 duplication and define more precisely the size of the duplicated segment we used a custom array CGH (aCGH) designed to encompass the Xq22 genomic region containing the PLP1 gene with maximum resolution. The aCGH profile confirmed the MLPA result and specified the breakpoint of the duplicated genomic segment covering from 102,761,000 to 103,513,000 (752 Kb of length) on Xq22 according to the GRCh37/hg19 assembly (Fig. 2a).

We next performed FISH analysis with a BAC probe specific to the PLP1 locus (RP11-832L2, Xq22.2) on lymphocyte metaphase spreads. Three distinct PLP1-specific signals were visualized consistent with the PLP1 duplication in a girl. Two probes bound to the normal Xq22 band of both X chromosomes together with an additional ectopic signal that bound just proximal to the CEB108/T7 1p36 subtelomeric probe (Fig. 2b). Another FISH analysis with BAC probes localized to X chromosome telomeres did not reveal a translocation involving the X chromosome. Hence, the additional copy of PLP1 results from an insertion into chromosome 1p36 rather than a classical translocation. A whole genome microarray confirmed the PLP1 duplication but did not reveal any copy number alteration in 1p36 region. MLPA analysis in the parents did not reveal any variation in PLP1 copy number. However, we wondered if this rearrangement was nevertheless inherited from a parental balanced rearrangement. Therefore we performed the PLP1 specific FISH analysis on the parents’ blood and did not detect any ectopic localization of the PLP1 gene. We thus conclude that this rearrangement was acquired either in the gametes of one of the parents or de novo.

Absence of skewed X inactivation pattern

A skewed X inactivation pattern resulting in inactivation of the X chromosome carrying the abnormal PLP1 gene has been suggested to explain why most females who carry PLP1 genomic alteration are asymptomatic [12]. We thus performed an analysis of the X inactivation pattern based on the extend HUMARA assay [8]. This analysis of DNA extracted from a blood sample from the patient did not reveal a significant skewed pattern for any of the four X chromosome loci explored (i.e. AR, PCSK1N, SLITRK4, and ZDHHC15). Although we cannot assert that there is not a skewed X chromosome inactivation pattern in the oligodendrocytes of the patient, the insertion of an extra-copy of PLP1 in an autosome did lead to a functional duplication irrespective of the X-inactivation pattern.

PLP1 over-expression in patient fibroblasts

We also quantified the PLP1 gene expression in fibroblasts from the patient and compared it to expression in normal control fibroblasts and fibroblasts from male patients with classical tandem PLP1 duplications. The duplicated controls and our patient showed a 1.88 and a 5.21-fold increase of the PLP1 expression respectively compared to the normal control. This result confirmed the functional duplication of the PLP1 gene leading to its overexpression in the patient’s fibroblasts.

Precise mapping of the breakpoint using genome walking strategy

To further characterize the insertion site of PLP1 on chromosome 1p36, we performed a genome walking strategy. This molecular procedure allows sequencing of the 1p36 telomeric flanking segment directly adjacent to the breakpoint of the duplicated genomic region of Xq22 (Fig. 2c). The Xq22 segment inserted exactly at the position chr1:140,332 (GRCh37/hg19) on chromosome 1p36. In addition, we identified 2 bp of microhomology shared by the 1p36 and the Xq22 boundaries genomic sequences.